Scribd
Upload
2K views4 pages

Image Classification in IDRISI and ILWIS

This document provides instructions for performing unsupervised and supervised image classification in IDRISI ANDES and ILWIS. For unsupervised classification in IDRISI ANDES, it describes using the cluster algorithm to automatically group pixels into a specified number of classes. For supervised classification, it outlines creating training data through digitization, extracting signatures, and using minimum distance classification. For ILWIS, it details density slicing, unsupervised clustering, and supervised maximum likelihood classification which involves creating sample sets for training data.

Uploaded by

Oluwafemi Opaleye
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOC, PDF, TXT or read online on Scribd
2K views4 pages

Image Classification in IDRISI and ILWIS

This document provides instructions for performing unsupervised and supervised image classification in IDRISI ANDES and ILWIS. For unsupervised classification in IDRISI ANDES, it describes using the cluster algorithm to automatically group pixels into a specified number of classes. For supervised classification, it outlines creating training data through digitization, extracting signatures, and using minimum distance classification. For ILWIS, it details density slicing, unsupervised clustering, and supervised maximum likelihood classification which involves creating sample sets for training data.

Uploaded by

Oluwafemi Opaleye
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOC, PDF, TXT or read online on Scribd
You are on page 1/ 4

Image Classification in IDRISI ANDES and ILWIS

(PRACTICAL)

Prepared by:
Oluwafemi. A. OPALEYE

For GPM III

Nov, 2010
IDRISI ANDES

Unsupervised Classification

1) Click Image Processing and navigate to Hard Classifier and to Cluster


2) Click the number of bands to be used and input them
3) Give name for the Output Image
4) Click Fine, under Generalization Rule
5) Set maximum number of clusters under the Clustering Rule
6) Click Ok to finish
7) To rename the classes; right click on individual class and click “update Legend”

Supervised Classification

1) Click Display Launcher under Display


2) Browse for the Composite Image and click ok to display it
3) Click on Digitize tool from the tool bar
4) Give name for the layer to be created and check polygon under layer type.
5) Click Ok to start digitizing
6) Digitize the identified land feature after zooming and right click to finish.
7) Pick the Save Digitized Data tool to save
8) Pick the Digitize tool to add other features to the active layer.
9) Click Image Processing; navigate to Signature and MAKSIG (for signature
extraction)
10) Input the vector file you created earlier
11) Click on Enter signature files names to name the Classes
12) Brose and input the bands
13) Click Ok to finish
14) Click Image Processing, navigate to Hard Classifier and then to MINDIST (for
minimum distance classification)
15) Check infinite under Maximum Search Distance
16) Browse for the Signature Files named earlier
17) Give output file name
18) Click next, then OK to Finish
ILWIS

Density Slicing

1) First Create a Group Domain


2) From File, navigate to Create then Domain
3) Under Type: check: Class and Group and click OK
4) From the Domain Group Menu; click Edit and navigate to Add Item
5) Give the Upper Bound value and Name. Note check from the band to be
“Sliced”
6) Do same for the other Group/Class
7) Do perform the “Slicing” click on Slicing from the Operation List or through
Operations-Image Processing – Slicing
8) Input the Raster Map
9) Give Name for the Output Raster Map
10) Under Domain; browse the Group Domain Created earlier and
11) Show to view the Raster Map

Unsupervised Classification

1) Goto Operations, navigate to Image Processing and then Cluster


2) Click on the number of bands to be used and insert them accordingly
3) Insert the number of Cluster
4) Give Name for the Output Raster Map
5) Click Show to view the result.

Supervised Classification (Maximum Likelihood)

1) Create Map List (File – Create – Map List),


2) Give name for Map List
3) Open the Map List (as Colour Composite)
4) Create Sample Sets (File – Create – Sample Set)
5) Specify name and also create a Domain
6) Choose ‘Map List’ create Maplist
7) Click OK
8) From the Domain Class Menu; click Edit and navigate to Add Item
9) Add the number of classes to be created
10) To create Sample Sets, choose the Normal Tool to digitize the identified Sites on
the image
11) Digitize as many as possible on the identified sites and click on exit editor when
through with digitizing.
12) Click on Operations – Image Processing – Classify
13) Choose “Maximum Likelihood”
14) Give Name for the Output Raster Map
15) Click show to view the result

You might also like